STS-MME algorithm for spectrum detection

Spectrum sensing is one of the key technologies in cognitive radio, and there are many methods for spectrum detection. The method based on eigenvalue detection does not need to understand the primary user signal or the noise power level. Most of the proposed methods require multiple receiving antennas or too many samples. In this paper, an efficient method of (primary user) PU signal detection based on eigenvalue is proposed. This method uses temporal smoothing and space smoothing technique to form a virtual multi group multi antenna structure, and then takes the mean of the covariance matrix to obtain the maximum and minimum eigenvalues for detecting the PU signal. The method is superior to other algorithms. First, it uses the power method to calculate the maximum and minimum eigenvalues, which reduces the computational complexity. Second, the use of a single antenna instead of multi-antenna can reduce system overhead. The most important thing is that the algorithm can effectively reduce the number of sample, improving the system real-time. The simulation results show that the system has good detection performance with small sample.